JP3463158B2 - Exciter for synchronous machine - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a synchronous machine excitation device, and more particularly to a technique for improving the stability of a power system.

[0002]

2. Description of the Related Art A conventional exciter for a synchronous machine has an active power source as disclosed in JP-A-61-280715 and JP-A-55-74400 for the purpose of improving the stability of a power system. , The generator frequency has been used as an input signal for stabilization. However, since the conventional method can only design optimized for a certain fixed power fluctuation frequency, it is not possible to secure sufficient stability when the system configuration and power flow conditions change significantly.

[0003]

SUMMARY OF THE INVENTION An object of the present invention is to speed up the fluctuation frequency and damping constant of a power system in real time,
An object of the present invention is to provide an exciting device for a synchronous machine, which can detect with high accuracy and is suitable for improving dynamic stability.

[0004]

In order to solve the above-mentioned problems, a means for obtaining a deviation between a voltage of a synchronous machine and a set value, active power change (ΔP) and frequency change (Δf) of the synchronous machine.
In the exciter for a synchronous machine, which has a power system stabilizing means for calculating a compensation output based on the above, and supplies an exciting current to the synchronous machine according to the output obtained by adding the compensation output to the deviation, the active power variation (ΔP ) Is divided into a component proportional to the frequency change (Δf) and its integral (∫Δfdt) respectively .
When a second proportionality constant (K1, K2) means for determining a first
And means for identifying the second proportionality constant, and the first identified
And the compensation output is automatically adjusted based on the second proportional constant.
Further, means for determining the active power variation each frequency variation of (ΔP) (Δf) and internal phase angle first and second proportionality constant for separating the component proportional to (Δδ) (K1, K2) , the A means for identifying the first and second proportionality constants is provided, and the compensation output is automatically adjusted based on the identified first and second proportionality constants. Here, the means for identifying the first and second proportionality constants are variables K1w, K from the active power change (ΔPi), the generator frequency change (Δfi), and the internal phase difference angle (Δδi).
2w and Dw are calculated, and variables K1w, K2w and Dw are
After removing the fluctuation component of the calculation result through the filter ,
By calculating K1 = K1w / Dw and K2 = K2w / Dw
First and second proportional constants (K1, K2) are obtained.

In order to improve the dynamic stability of the synchronous machine, it is necessary to know what the operating state of the synchronous machine including the power system is. The change ΔP of the synchronous machine active power is directly related to the dynamic stability, and in order to improve the stability, the fluctuation frequency and the damping constant of ΔP during operation and the phase relationship between each physical quantity are real-time. Need to know. Therefore, in the present invention, the variation ΔP of the active power is set.
Of the generator frequency is separated into proportional constants K1 and K2 which are components proportional to the change Δf of the generator frequency and the integral ∫Δfdt, and the power stabilizing means (PSS) is based on the detected values of K1 and K2.
Parameter (P) is automatically adjusted to improve the stability of the synchronous machine. As the first means for detecting K1 and K2,
First, the active power Pi and the generator frequency fi are differentiated, and the active power change ΔPi, the generator frequency change Δfi,
Further, the integral ∫Δfdt is obtained (here, the subscript i indicates the value at the sampling time i). Next, ΔP
i and K1 * Δfi + K2 * ∫Δfidt squared J is obtained by the least squares method, and ∂J / ∂K1 = 0, ∂J / ∂K
From 2 = 0, K1 and K2 that minimize the error J are obtained. K
As a second means for detecting 1, K2, first, the change ΔPi in the obtained active power and the change Δf in the generator frequency are calculated.
i, coefficients A11, A12, A2 from the internal phase difference angle Δδi
1, A22, B1, B2 are obtained, and intermediate outputs Dw, K1w, K2w are obtained from these coefficients. Next, K1
= K1w / Dw, K2 = K2w / Dw
1, K2 is detected. Thus, ΔPi can be expressed as ΔPi≈K
1Δfi + K2∫Δfidt, or ΔPi≈K1Δ
By separating into fi + K2Δδi, the state of ΔPi is identified, and from these, it is necessary for adaptive control of the power system stabilizing means (PSS) such as the power fluctuation frequency of the synchronous machine, the damping time constant, and the phase difference between the control signals. Information can be identified in real time.

[0006]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is an exciting device for a synchronous machine showing an embodiment of the present invention. In FIG. 1, 1 is a synchronous machine, 2
Is an instrument transformer (PT), 3 is an instrument current transformer (CT),
4 is a signal detection circuit, 5 is an AVR voltage setting value, 6 is a subtraction circuit, 7 is an addition circuit, 8 is a gate pulse generator (GP).
G), 9 is an integrating circuit, 10 is a K1, K2 detecting circuit, 11
Is the K1 set value, 12 is the power system stabilizer (PSS
(P)), 13 is an excitation transformer (EXTR), 14 is a thyristor circuit (THY), and 15 is a field breaker (FCB).
Represents In the present embodiment, the terminal voltage of the generator 1 is supplied to the signal transformer 4 via the instrument transformer 2 and the armature current via the instrument transformer 3. The signal detector 4 outputs the generator voltage Vg necessary for the excitation control, the active power change ΔP, and the frequency change Δf. Vg and AVR voltage set value V by subtraction circuit 6
The deviation of kg is taken and this deviation is output to the addition circuit 7.
Further, the integrator 9 outputs the time integral value ∫Δfdt of Δf. Next, in the K1 and K2 detection circuits, ΔP, Δf, ∫Δ
fdt is input and ΔP is set to Δf by the K1 and K2 detection circuits.
And proportional coefficients K1 and K2 when linearly approximated by ∫Δfdt are obtained. K1 is compared with the K1 set value K1ref11, and the power system stabilizer (P
The parameter P of SS) 12 is controlled. The same applies to the proportionality coefficient K2. The compensation output of the power system stabilizing device (PSS) 12 obtained based on the parameter P is output to the adding circuit 7. The compensation output of the power system stabilizing device (PSS) 12 is added to the deviation between Vg and Vkg, and a gate pulse is generated from the gate pulse generating device 8, and the thyristor circuit 1
Turn 4 on and off. Turning on this thyristor circuit 14,
When turned off, the exciting current is supplied from the exciting transformer 13 to the field winding of the generator 1 via the field breaker 15. As described above, in the present embodiment, the components proportional to ΔP and Δf and ∫Δfdt, that is, the proportional coefficients K1 and K2, are obtained and compared with the set values, respectively, and the power stabilization device (PSS) 12 according to these deviations is calculated. The stability of the synchronous machine is improved by automatically adjusting the parameter (P).

FIG. 2 shows details of the K1 and K2 detection circuits. The active power Pi and the generator frequency fi are detected using the generator voltage and the armature current detected by PT2 and CT3, and these signals are further changed by the incomplete differentiation circuit 16 with respect to the respective time changes ΔPi and Δfi. Ask for. Next, the integrator 9 detects the time integration value ∫Δfidt of Δfi. Using these signals, the J detection circuit 17 outputs ΔP
An error J when i is linearly approximated by Δfi and ∫Δfidt is obtained by the least square method of the equation (1).

[Equation 1] Then, by the J minimization circuit 18, ∂J / ∂K1 = 0,
From ∂J / ∂K2 = 0, K1 and K2 that minimize the error J are obtained. The subscripts of ΔPi and Δfi indicate values at time i sampling. Specific calculation method (2)
It is shown in equation (7).

[Equation 2]

[Equation 3] Find the condition that minimizes J.

[Equation 4]

[Equation 5] Looking at K1 and K2 as variables,

[Equation 6] here,

[Equation 7] far.

FIG. 3 shows the details of the K1 and K2 detection circuits shown in FIG. FIG. 3 shows ΔPi, Δfi, and ∫Δfidt.
The low-pass filter 21 having the same frequency characteristic is added to the detection signal of 1), the influence of the waveform distortion such as the detection circuit noise and the system disturbance is removed, and then the signal is input to the J detection circuit. In the K1 and K2 detection circuit of FIG. 3, since the low-pass filter 21 is added, signals such as ΔPi and Δfi are delayed, but since the low-pass filter 21 has the same frequency characteristic, ΔPi, Δfi, and ∫Δf
The mutual relationship of idt maintains a constant ratio, and the proportional coefficients K1 and K2 that minimize J are not so much influenced by the low-pass filter 21. Therefore, by adding the low-pass filter 21, it is possible to remove the influence of noise and waveform distortion without delaying the signal detection of K1 and K2. The internal phase difference angle Δδi may be used instead of ∫Δfidt.

FIG. 4 shows a K1 and K2 detection circuit which has a response almost equal to that of the K1 and K2 detection circuit shown in FIG. 3 and which is stronger against the influence of noise and waveform distortion. The K1 and K2 detection circuits in FIG. 4 have coefficients A11, A12, A21, A having an arithmetic circuit 20 and a low-pass filter circuit 21.
22, B1, B2 detection circuit 19 and Dw, K1w, K2
The w arithmetic circuit 22 and the K1 and K2 arithmetic circuits having the low-pass filter circuit 23 and the division circuit 24 are included. First,
For simplification, Δδi = ∫Δfid is set. Arithmetic circuit 20
At Δfi and Δδi, the coefficient A which is the sum of these products
11, A12, A21, A22, B1, B2 ((7)
Formula). These are input to the low-pass filter 21 having the same frequency characteristic to remove the influence of noise and waveform distortion of the input signal. Up to this point, the method is the same as in FIG.
Outputs A11, A12, ..., B of the low-pass filter 21
Instead of directly detecting K1 and K2 using 1 and B2, the intermediate outputs Dw, K1w, and K2w are once calculated.

## EQU00008 ## Dw = A11.A22-A12.A21 K1w = A22.B1-A12.B2 (8) K2w = -A21.B1 + A11.B2 These Dw, K1w, and K2w change through the low-pass filter 23. After further removing the components, K1 = K1
K1 and K2 are detected by w / Dw, K2 = K2w / Dw. Here, the detection values of the denominator are A11, A12, A
Although the determinant when viewed as a matrix having 21 and A22 as components is used, any variable may be used as long as it is separable so that the denominator and the numerator are also changed. According to the numerical simulation, Dw, K1w, and K2w respond with substantially the same ratio, so that even if the low-pass filter 23 is used, there is almost no detection delay of K1 and K2.

K1, K in FIGS. 2, 3 and 4 described above
2 detection circuit, K1, K required for power system stabilization
2 can be identified, but when the amount of change in the signals of ΔPi, Δfi, and Δδi used to identify K1 and K2 is small, the numerical calculation becomes unstable. For this reason,
Coefficients A11, A12, ... For calculating K1 and K2
It is necessary to monitor the absolute values of B1 and B2.

First, in FIG. 5, the coefficients A11, A12, ...
The monitoring circuit of the absolute value of B1 and B2 is shown. FIG. 5 shows the coefficient A
An AND condition in which the absolute values of 11, A12, ..., B1, B2 are all equal to or greater than a constant value ε is detected, and only when this condition is satisfied, that is, when IDENTON = 1, the update calculation of K1 and K2 is performed. To do. By this means, K1 and K2 can be identified without being numerically unstable.

FIG. 6 shows the coefficients A11, A12, A13, A.
22 shows a circuit for monitoring the absolute value of the determinant Dw of 22. Figure 6
Only when the absolute value of Dw, which is the denominator of the division operation, is equal to or greater than a constant value ε (in FIG. 4, since Dw is the denominator, Dw = 0
In order to avoid), the update calculation of), K1, and K2 is performed. By this means, K1 and K2 can be identified without being numerically unstable even in the circuit of FIG.

Next, FIG. 7 shows a method of obtaining the fluctuation frequency of the active power of the synchronous machine using K1 and K2 thus obtained. The motion of the synchronous machine can be approximated by a secondary vibration system. Here, the change amount Δω of the shaft speed of the synchronous machine is equal to the change amount Δf of the generator frequency, and the internal phase difference angle Δ
δ is equal to the integral of Δf ∫Δfdt. Therefore,

## EQU9 ## ΔP≈K1Δf + K2∫Δfdt That is, ΔP≈K1Δω + K2Δδ (9), the secondary vibration system of FIG. 7 is obtained. From FIG. 7, the power fluctuation frequency ωs and the damping time constants α, α1, and α2 are set to (10) to (1
It can be identified using the equation (5).

[Equation 10]

[Equation 11]

[Equation 12]

[Equation 13]

[Equation 14]

[Equation 15]

In the above embodiment,

[Equation 16] ΔPi = K1Δfi + K2∫Δfidt (16) is approximated, but when the power system can be represented as the generator-infinity system shown in FIG. 9, the internal phase difference angle Δδ is calculated from the vector diagram of FIG. You can ask. In FIG. 9, Eq is the q-axis rear voltage of the synchronous machine, VG is the terminal voltage of the synchronous machine, VB is the infinite bus voltage, Xq is the q-axis synchronous reactance of the synchronous machine, Xe is the system reactance, P is the active power, and Q is Represents the reactive power, and in FIG. 8, δ is the phase difference angle between VB and Eq,
δq represents a phase difference angle between VG and Eq, and δ _B represents a phase difference angle between VG and VB. The vector relation is

[Equation 17]

[Equation 18] from now on,

[Formula 19]

[Equation 20] Therefore,

[Expression 21] δ = δq−δ _B (21) Using this δ,

## EQU22 ## By approximating ΔP≈K1Δfi + K2Δδi (22), K1 and K2 can be obtained. Here, the subscript i of Δδi indicates the i-th data of Δδ.

In the ideal secondary vibration system shown in FIG. 7,

[Expression 23] ΔP = K1Δf + K2Δδ (23) However, in FIG. 7, ΔPm (mechanical torque change) has a slower response than the electrical torque change ΔP, and therefore can be regarded as ΔPm = 0. The equation of motion of a synchronous machine that operates continuously with a general electric power system does not exactly match the ideal secondary vibration system as shown in FIG. 7, but the model of FIG. 7 can be applied approximately. Therefore, the active power ΔP

By identifying K1 and K2 from the measured data ΔP, Δf, and Δδ as ΔP≈K1Δf + K2Δδ (24), the power fluctuation frequency ωs and the damping time constants α, α1, and α2, which are indexes of the power system stability, are identified. It is possible to detect the information necessary for the stability improvement control of.

The present invention uses the power fluctuation information ΔP, Δ
K1 and K2 are identified online using f, Δδ, and the like. FIG. 10 shows a K1 and K2 detection circuit using Δδi. In FIG. 10, since Pi, fi, and δi can all be electrically detected, the differentiating circuit 16 need only be one stage (two stages in FIGS. 2 and 3), and the integrating circuit 9 is not necessary (see FIG. 2). , Δi is used instead of ∫Δfdt in FIG. 3.)
Therefore, the circuit configuration can be simplified.

Next, when the power fluctuation between the systems becomes a problem, the power fluctuation frequency of the synchronous machine becomes a plurality of modes. in this case,

When approximated as ΔP≈K1Δf + K2Δδ (25), these modes are averaged values for K1 and K2, and the actual fluctuation cannot be expressed correctly. Practically about 1.0 Hz generated between nearby synchronous machines
(0.5-1.5Hz) local shaking and about 0.3H
It suffices to consider the system fluctuation of z (0.1 to 0.3 Hz). For this signal separation, a bandpass filter may be prepared for local oscillation and intersystem oscillation, and two sets of circuits (FIG. 11) in which the passing frequency is divided into local oscillation frequency and intersystem oscillation frequency may be prepared. In FIG. 11, Pi, fi
Are input to the bandpass filter 26, respectively, and ΔP
i, Δfi, and Δδi through the integrator 9 are J, K1, and K2.
It is input to the detection circuit 10. The power fluctuation frequency detection circuit 27 detects the fluctuation frequency fs based on the detected K1 and K2. The first-order delay circuit time constants T1 and T2 are tuned so that the center frequency fo of the bandpass filter 26 matches the detected oscillation frequency fs. That is, for example, it is assumed that the first-order lag circuit time constants T1 and T2 in FIG. 11 are tuned to a local oscillation of 1 Hz (T1 = 2).
π × 2.0, T2 = 2π × 0.667). In this case, even if an intersystem frequency of 0.1 to 0.3 Hz is applied to the circuit of FIG. 11, these signals are attenuated by the bandpass filter 26, so the obtained K1 and K2 are the local fluctuations of 0. It can be set to a value reflecting 5 to 1.5 Hz. Further, by tuning the first-order lag circuit time constants T1 and T2 so that the center frequency fo of the bandpass filter 26 of FIG. 11 matches the detected fluctuation frequency fs, the local fluctuation frequency can be detected more accurately. . The center frequency fo of the bandpass filter 26 of FIG.

Since fo = (1 / 2π) (1 / √T1T2) (26), the bandwidth of T1 and T2, for example, T1 = K ² T2 (K ≧ 1),

## EQU27 ## fo = (1 / 2π) (1 / T1) (1 / K) (27), and T1 is adjusted so that fo matches fs.
Therefore, by automatically tuning fo so that it matches fs, it is possible to accurately obtain K1 and K2 in the target area even when local fluctuation and intersystem fluctuation are included.

FIG. 12 shows the frequency characteristic of one stage of the bandpass filter used in FIG. 1 / T1 and 1 / T2 are polygonal frequencies on the low frequency side and the high frequency side, and ωo =
At 1 / √T1T2 or fo = (1 / 2π) ωo, the gain becomes maximum and the phase change becomes zero.

[0019]

As described above, according to the present invention,
Power system fluctuation information such as active power and generator frequency is used as an input signal, and active power is decomposed into proportional components of the generator frequency and the time integral value of the generator frequency, and the identified proportional coefficient K is obtained.
The fluctuation frequency and the damping constant of the power system can be detected in real time from 1 and K2 with high speed and high accuracy, and the system stability can be significantly improved regardless of the operating state of the power system. In addition, when the absolute values of the active power change and the frequency change are small, the proportional constants K1 and K2 are updated only when the active power change and the frequency change are equal to or more than a certain value. Without becoming K1, K2
Can be identified. Further, by adding a low-pass filter to the K1 and K2 detection circuits, it is possible to remove the effects of noise and waveform distortion. Also, when there are multiple power modes, a bandpass filter is provided to separate these modes, and the target frequency is adjusted by automatically tuning the center frequency of the bandpass filter to match the power fluctuation frequency. The proportional constants (K1, K2) of the shaking frequency can be accurately obtained.

[Brief description of drawings]

FIG. 1 is an excitation device for a synchronous machine showing an embodiment of the present invention.

FIG. 2 is a detailed diagram of a K1 and K2 detection circuit according to the present invention.

FIG. 3 is a detailed diagram of a K1 and K2 detection circuit according to the present invention.

FIG. 4 is a detailed diagram of a K1 and K2 detection circuit according to the present invention.

FIG. 5 shows coefficients A11, A12, ..., B1, B2 of the present invention.
Absolute value monitoring circuit

FIG. 6 shows coefficients A11, A12, A13, A22 in FIG.
Circuit for absolute value of determinant Dw of

FIG. 7: Ideal secondary vibration system

FIG. 8 Vector diagram of synchronous machine

[Figure 9] One machine infinity system diagram

FIG. 10 is a K1 and K2 detection circuit using Δδ as an input according to the present invention.

FIG. 11 is a K1 and K2 identification circuit for automatically tuning to a power fluctuation frequency according to the present invention.

FIG. 12 is a bandpass filter frequency characteristic diagram.

[Explanation of symbols]

1 ... Synchronous machine, 2 ... Transformer for instrument, 3 ... Current transformer for instrument, 4
... signal detection circuit, 5 ... AVR voltage setting device, 6 ... subtraction circuit, 7 ... addition circuit, 8 ... gate pulse generator, 9 ... integration circuit, 10 ... K1, K2 detection circuit, 11 ... K1 set value circuit, 12 … Power system stabilizer, 13… Excitation transformer, 14… Thyristor circuit, 15… Field interruption (FCB:
Field Circuit Breaker), 16 ... Incomplete differentiation circuit, 17
... J detection circuit, 18 ... K1, K2 detection circuit, 19 ... A1
1, ..., B1, B2 detection circuit, 20 ... A11F, ..., B
1, B2 arithmetic circuit, 21 ... Low-pass filter circuit, 22
... Dω, K1ω, K2ω arithmetic circuit, 23 ... Low-pass filter circuit, 24 ... Division circuit, 25 ... K1, K2 arithmetic circuit, 26 ... Bandpass filter, 27 ... Power fluctuation frequency detection circuit

─────────────────────────────────────────────────── ─── Continuation of the front page (58) Fields surveyed (Int.Cl. ⁷ , DB name) H02J 3/00-5/00 H02P 9/14

Claims (6)

(57) [Claims] 1. A means for obtaining a deviation between a voltage of a synchronous machine and a set value, and a power system stabilizing means for calculating a compensation output based on a change amount (ΔP) and a frequency change amount (Δf) of the active power of the synchronous machine. In a synchronous machine exciter that supplies an exciting current to a synchronous machine according to an output obtained by adding a compensation output to the deviation, the active power change (ΔP) is the frequency change (Δf) and its integration, respectively. The means for determining the first and second proportional constants (K1, K2) to be separated into components proportional to (∫Δfdt) and the means for identifying the first and second proportional constants are provided, and the identified first An exciting device for a synchronous machine, characterized in that the compensation output is automatically adjusted based on a second constant of proportionality. 2. A means for obtaining a deviation between the voltage of the synchronous machine and a set value, and a power system stabilizing means for calculating a compensation output based on the active power change (ΔP) and the frequency change (Δf) of the synchronous machine. In the exciter for a synchronous machine, which supplies an exciting current to a synchronous machine according to an output obtained by adding a compensation output to the deviation, the active power variation (ΔP) is respectively the frequency variation (Δf) and the internal phase difference. Means for obtaining first and second proportional constants (K1, K2) that separate into components proportional to the angle (Δδ) and means for identifying the first and second proportional constants are provided, and the identified first An exciting device for a synchronous machine, characterized in that the compensation output is automatically adjusted based on a second constant of proportionality. 3. The means for identifying the first and second ratio constants according to claim 2 , wherein the active power variation (ΔP
i), the frequency variation ([Delta] fi), the internal phase angle (Δδi) from the variable K1w, K2w, the Dw determined, the
Variables K1w, K2w and Dw are passed through a low pass filter
After removing the fluctuation component of the calculation result , K1 = K1w / D
An exciter for a synchronous machine, characterized in that the first and second proportional constants (K1, K2) are obtained by calculating w, K2 = K2w / Dw. 4. The absolute value of the active power variation (ΔP) and the frequency variation (Δf) according to claim 3 ,
When the first and second proportional constants (K1, K2) are calculated so small as to cause instability in numerical operation, in order to avoid this instability, only when the absolute value of the variable Dw is a certain value or more,
An exciting device for a synchronous machine, characterized in that the first and second proportional constants (K1, K2) are updated. 5. The fluctuation frequency and damping constant of active power of a synchronous machine according to claim 1 , wherein the determined first and second proportional constants (K1, K2) are used to identify the fluctuation frequency and the damping constant. An exciter for a synchronous machine, comprising: 6. The method according to any one of claims 1 to 5 , wherein when there are a plurality of power modes, a bandpass filter for separating these modes is provided, and a signal after the mode separation is divided into active power variation components. (ΔP) and frequency variation (Δf), and means for automatically tracking the center frequency of the bandpass filter so as to match the fluctuation frequency of the active power of the synchronous machine are provided . An exciting device for a synchronous machine, characterized by identifying 1 and a second proportionality constant (K1, K2).